Nanotribology of Poly (Dimethylsiloxane) Melt Confined

Tribology Online, 1, 2 (2006) 29-33. ISSN 1881-2198 DOI 10.2474/trol.1.29

Nanotribology of Poly(dimethylsiloxane) Melt Confined between Hydrophobic Surfaces

Shinji Yamada*

Analytical Research Center, Kao Corporation 2606 Akabane, Ichikaimachi, Haga, Tochigi 321-3497, Japan *Corresponding author: [email protected]

( Manuscript received 24 August 2006; accepted 10 October 2006; published 31 October 2006 )

Friction measurements were carried out for a poly(dimethylsiloxane) (PDMS) melt (Mw ≈ 80000) confined between hydrophobic surfaces using the surface forces apparatus. The PDMS films were prepared by two different procedures: i) compression of a droplet into a hard-wall state (compressed system); ii) adhesive contact of two thin films cast on each substrate from solution (cast system). The dynamic thicknesses were 1.4 nm for the compressed system and 2.0 nm for the cast system. Despite the large thickness, the friction of the cast system was larger than that of the compressed system. Large thicknesses generally give low friction; the unusual result suggests that the confined structures may be different between the two systems. The PDMS molecules in both systems lay parallel to surfaces, but the extent of ordering could be much higher for the compressed system. The compressed film has a layer structure and slipping mainly occurs between the layers, resulting in the low friction. On the contrary, the cast system should have a disordered structure; molecules may interdigitate to each other and possibly form bridges across the sliding surfaces, which could induce large friction. The effect of film the preparation procedures on molecular ordering is discussed.

Keywords: surface forces apparatus, nanotribology, confinement, poly(dimethylsiloxane), molecular layering, squeeze flow

(PDMS) melt using the SFA. The molecular weight of 1. Introduction the PDMS melt was about 80000, which is large enough to have chain entanglement effect (an entanglement The structures and dynamics of thin polymer melt 10 ) limit of PDMS is 12000) . However, the results films confined between two surfaces are of importance showed that the dynamic thicknesses equal to two, three in many academic research areas as well as for several or four molecular layers, which are an order of practical applications such as lubrication and polymer magnitude thinner than the typical size of the random processing. Due to the ability to measure the dynamic coil conformation. Measured friction properties also thickness and shear force of an intervening film suggested that the extent of ordering of PDMS accurately during sliding, the surface forces apparatus molecules in the films is exceptionally high; shear is (SFA) technique has been employed to study the mainly accomplished by the slipping between the dynamic behavior 1-4). One of the characteristic layered planes. However, the mechanism underlying this observations of confined polymer melts compared with exceptionally high molecular ordering (layering) of the simple liquid lubricants is relatively large dynamic polymer chains is not clear. thicknesses5,6). When confined between two surfaces by In this paper, we describe the tribological normal compression, polymer melts often produce experiment of the thin PDMS films, prepared by two nonequilibrium repulsions such as steric forces and/or different procedures, sheared between two hydrophobic viscosity forces, which tend to prevent the thickness surfaces. The nanometer-thick films of PDMS melt were decrease 7). If the polymer melts have large chain cast on hydrophobic substrates from solution, and the entanglement effects, the nonequilibrium interactions two cast films were made in adhesive contact. The are sometimes large enough for the films to stabilize friction of the film was measured using the SFA, and the against compression at thicknesses roughly comparable results were compared with those of a hard-wall PDMS to the size of the random coil dimensions. film prepared by the compression of a bulk droplet. The In a previous study8,9), we investigated the structures effect of film preparation procedures on the molecular and dynamics of a confined poly(dimethylsiloxane)

ordering (layering) is discussed, which gives us a new The compressed PDMS film was obtained by the insight to obtain highly-ordered layer structures for long following procedure. The two DDAB-coated mica chain polymer melts in confinement. surfaces were positioned in a crossed cylinder configuration in the measuring chamber of a SFA3 2. Experimental Methods (SurForce Corp., USA)13). The droplet of PDMS melt (vol. ∼ 0.1ml) was put on the DDAB-coated surface The poly(dimethylsiloxane) (PDMS) melt using a thin syringe needle and then the chamber was investigated is a commercial silicone oil obtained from purged with dry nitrogen gas for 12 h. Some P O was Toray Dow Corning Silicone Co., Japan and was used as 2 5 also placed inside the sealed chamber to keep the received8). The weight-average molecular weight M is w internal atmosphere completely dry at all times. The two about 80000 and the estimated polydispersity of the surfaces were approached stepwise by decreasing sample is 1.4. The size of the random coil conformation surface distance (applying an normal pressure P). The is estimated to be about 10 nm9), and the width of the interface was allowed to equilibrate at each step until PDMS chain is about 0.7 nm11). the surface distance (PDMS thickness) became stable. The hydrophobic surface used in this study was a Finally, the surfaces become flat because of the elastic self-assembled monolayer of double-chained surfactant deformation of the glue layer under each mica substrate DDAB (Didodecyldimethylammonium bromide, as shown in Figure 1a. A hard-wall film was attained at ACROS Organics, USA) on cleaved mica8, 12 ). The P = 3.0 MPa. Before applying sliding, the hard-wall molecular area was about 0.50 nm2 per molecule, which film was left to equilibrate for 1 h. corresponds to an almost close-packed monolayer for The cast PDMS system was prepared as follows this surfactant. The surface energy was 25 mJ/m2 for the (Figure 1b). The PDMS was dissolved in hexane at monolayer8,12). concentration 0.03 mg/mL. The solution (vol. 10 µL) was dropped and spread onto the DDAB-coated mica a) Compressed system 2 DDAB-coated mica sheet (about 10 × 10 mm ) and left for solvent evaporation. a The estimated thickness of the PDMS P film was about 3 ∼ 5 nm, and the RMS roughness was less than 0.1 nm (obtained over an area of 3 × 3 µm using atomic force microscope). The two PDMS films cast on mica were installed into the SFA chamber and compression and hard-wall state then the chamber was purged with dry nitrogen gas for squeeze out 12 h and some P2O5 was also placed. When the two b) Cast system PDMS surfaces approached slowly, they jumped into adhesive contact at the surface distance about 20 nm. Cast The contact thickness decreased gradually and reached films equilibrium in 1.5 ∼ 2 h. Friction measurements were made at the adhesive contact condition (P = 0). The SFA was equipped with a bimorph slider that generates lateral motions at constant velocity V (0.001 approach in air adhesive contact to 0.2 µm/s) and resulting friction force F was measured by means of a friction device14). The dynamic thickness Fig.1 Two different procedures to prepare molecularly (thickness during sliding) D and real contact area A thin PDMS films between hydrophobic surfaces. were directly measured using multiple beam interferometry (MBI) 15 ). The friction force was Friction normalized by the contact area and the shear stress S (= Load, L Force, F F / A) was evaluated. The schematic illustration of the (Pressure, P)

a PDMS Contact Area, A Prior to the PDMS experiment, the influence of hexane on Film the structure and properties of the DDAB monolayer was 0.7 nm examined. The droplets of blank hexane solution were spread on the DDAB surfaces and left for complete solvent DDAB evaporation, and then SFA friction measurements were made Film Thickness, D Layer for the two DDAB monolayers in adhesive contact. The result for the DDAB film before the hexane treatment is shown in Mica Substrate Figure 3. The friction force (shear stress) of the DDAB film Sliding before and after the hexane treatment was not different within Velocity, V the experimental error range (± 10 %), and the friction trace Fig.2 Schematic drawing of the contact region in the pattern was basically the same before and after the treatment. Further, AFM analysis did not show any structural changes of SFA friction measurements. the DDAB surfaces by the treatment.

Japanese Society of Tribologists (http://www.tribology.jp/) Tribology Online, Vol. 1, No. 2 (2006) / 30 Nanotribology of Poly(dimethylsiloxane) Melt Confined between Hydrophobic Surfaces

sliding interface is shown in Figure 2. All the results the adhesive pressure was not very different between the reported in this paper were obtained with molecularly two PDMS systems. This analysis is also inconsistent smooth hydrophobic substrates (wearless friction). The with the large shear stress of the cast system.b experimental room was kept at a fixed temperature of 23°C ± 0.2°C. 106

3. Results and Discussion [Pa] Cast PDMS S The dynamic thicknesses of the confined PDMS melt were 1.4 nm for the compressed system and 2.0 nm Compressed for the cast system (Table 1). Comparison with the 105 PDMS width of the polymer backbone implies that the PDMS molecules in the both systems lay parallel to substrate DDAB monolayers surfaces; the dynamic thicknesses correspond to two Shear Stress, molecular layer thick for the compressed system and -3 -2 -1 0 three molecular layer thick for the cast system. For the 10 10 10 10 compressed system, the dynamic thickness equaled the Sliding Velocity, V [µm/s] static hard-wall thickness (thickness before sliding). For the cast system, the static thicknesses did not always Fig.3 Shear stresses as a function of sliding velocity equal the dynamic thickness, larger thickness (D = 2.8 for the two PDMS systems. The shear stress for nm) was obtained at some contact positions. However, the DDAB/DDAB interface is also shown for the thickness decreased to 2.0 nm when sliding was comparison. Sliding conditions: D = 2.0 nm and applied. We should note that all the static hard-wall P = 0 MPa (adhesive contact) for the cast thicknesses and the dynamic thicknesses for the two PDMS, D = 1.4 nm and P = 3.0 MPa for the PDMS systems obtained here roughly equaled the compressed PDMS, and P = 0 MPa (adhesive integral multiples of the width of the PDMS backbone. contact) for the DDAB monolayers. The dependence of the shear stress on sliding velocity for the PDMS systems is shown in Figure 3. The shear stress for the two DDAB monolayers In Figure 3, we did not add error bars for the results (hydrophobic substrates) in adhesive contact (P = 0 of the compressed system, because it was difficult to MPa) is also shown for comparison. The shear stress for keep applied load and pressure the same at different the cast system was larger than that of the compressed contact positions. The typical reproducibility of the system, despite the large dynamic thickness. Generally, friction measurements for the system at different contact large dynamic thicknesses give low friction for most of positions was within the range of ± 15 %. the confined fluids; the result obtained here is opposed The relationship between shear stress S and sliding to the common behaviors. velocity V can be fit by the following equation: S ∝ V α (1) Table 1 Results of the friction measurements for the where α is a constant. The α values for the two PDMS two PDMS systems. systems are also included in Table 1. The α parameter indicates the shear mechanisms whether “solidlike” Applied Dynamic α friction or “liquidlike” rheological flow are observed: pressure, P thickness, for ideal friciton α = 0 and for Newtonian flow α = 4,6,16) [MPa] D [nm] 1 . The α for the two PDMS systems are the same Compressed and rather close to the ideal value of solidlike friction. 3.0 1.4 0.18±0.01 system Therefore, the microscopic shear mechanisms of the two 0 (adhesive films are not very different and mainly governed by the Cast system 2.0 0.18±0.02 contact) slipping between molecules. Figure 4 shows the possible confined structures of the two PDMS films. According to the results described 8) When we discuss the effect of applied pressure on above and in our previous paper , we estimate that the shear stress in Figure 3, the following consideration compressed PDMS system has a highly-ordered layer should be made. Although the applied pressure was zero for the cast system, self-adhesion produces attractive b For both systems, we could not perform the friction pressure at the contact interface. A simple estimation of measurements under large applied pressure at different sliding the adhesion strength using the Laplace pressure velocities, because the contact interfaces damaged easily equation gives ∼ 2.5 MPa for the cast system. Therefore, during sliding (the DDAB layers came off from mica surfaces and formed wear particles).

Japanese Society of Tribologists (http://www.tribology.jp/) Tribology Online Vol. 1, No. 2 (2006) / 31 Shinji Yamada

structure (Figure 4a). The small hard-wall thickness squeeze flow (under a large shear rate). Recent X-ray (two molecular layers) suggests that molecules in the scattering experiment supports this mechanism, film are easily expelled from contact interface during shear-induced layering is observed for a PDMS melt at a squeezing, which implies the low interdigitation of liquid/solid interface19). In addition, boundary slip may molecules between layers. In addition, low friction also occur at the hydrophobic DDAB surfaces 20 ), which fits the expectation of the high ordering because of the could also contribute to the layering. When the easy slippage between ordered molecular planes8,9). In thickness of the PDMS melt reaches the hard-wall state contrast, the relatively large dynamic thickness of the by compression, the extent of the molecular layering on cast system suggests that the molecules are less substrate should be already exceptionally high. expellable than the compressed system, probably due to The squeeze out of molecules is also expected for the molecular entanglement effect. Large friction for the the cast system when the two cast films are in adhesive cast system compared with the compressed system contact. The thickness gradually decreases with time implies the interdigitation between molecules and and reaches the final thickness by squeezing. However, molecular bridges between two sliding surfaces (Figure the effect is not large enough for the long PDMS chains 4b)17,18). If lateral sliding motions play a major role to to order into layers. The molecular conformation in the enhance the molecular ordering, the cast film (three system should be mainly governed during the casting of layer thick) should have transited into two layers during the nanometer thick film on substrate from solution. The shear. However, the layering transition did not occur but random coil conformation in the solution becomes surface damage occurred instead. This analysis leads to thinner during the solvent evaporation, and molecular a conclusion that not the lateral sliding motions but the chains lay parallel to surfaces due to the interaction normal compression (squeeze out) is important to between molecules and surfaces21). However, there is no enhance the molecular ordering in confined PDMS driving force for the molecules to disentangle the PDMS films. chains with each other; the extent of the ordering is low. This is consistent with the recent synchrotron X-ray 22) a) Compressed System b) Cast System reflectivity study of PDMS melts on surface . When the two cast PDMS films are made in contact, the molecules at interface interdigitate each other. Especially, the chain ends could have a large effect on the interpenetration and restructuring the surface segments23,24). Contact pressure squeezes some amount of molecules as was mentioned, and the final film has an interdigitated (entangled) structure. Restructuring may induce molecular bridges between the two surfaces, Fig.4 Schematic illustrations of the possible confined which is presumably the major cause of high friction. structures in the two PDMS systems. The Lateral sliding motions cannot rearrange the molecules in the both films lay parallel to interdigitated film structures into well-ordered layers substrate surfaces, but the molecular but induce surface damage. conformations may be different. a) The In summary, we have investigated how long PDMS molecules in the compressed system have rather chains form layer structures in confinement using the a flat conformation and molecules belong to a SFA. The results imply that the pressure-driven squeeze specific layer. b) The molecules in the cast flow of PDMS molecules at large thickness is important system could interdigitate between each other to form a well-ordered layer structure, which gives us a and molecular bridges may be formed between new insight to manipulate large molecules under the two opposed surfaces. confinement.

4. Acknowledgements Now we discuss how the film preparation procedures make the difference of the film structures. If The author is grateful to Dr. Kazuo Maki for the estimated models shown in Figure 4 are adequate, enlightening discussions, and to Kao Corporation for the major factor that determines the extent of ordering is permission to publish this paper. the compression (squeeze out) of the PDMS molecules at large separations (large thicknesses). We speculate that the pressure-driven squeeze flow of molecules, which has a large shear rate, contributes to the flat 5. References molecular conformation and layering for the compressed system. Because of the large flexibility of [1] Israelachvili, J., McGuggan, P. M. and Homola, A. siloxane backbone, the PDMS molecules at surfaces M., “Dynamic Properties of Molecularly Thin may be disentangled and flattened along with the Liquid Films,” Science, 240, 4849, 1988, 189-191.

Japanese Society of Tribologists (http://www.tribology.jp/) Tribology Online, Vol. 1, No. 2 (2006) / 32 Nanotribology of Poly(dimethylsiloxane) Melt Confined between Hydrophobic Surfaces

[2] Granick, S., “Motions and Relaxations of [14] Luengo, G., Schmitt, F. J., Hill, R. and Confined Liquids,” Science, 253, 5026, 1991, Israelachvili, J. N., “Thin Film Rheology and 1374-1379. Tribology of Confined Polymer Melts: Contrasts [3] Bhushan, B., Israelachvili, J. and Landman, U., with Bulk Properties,” Macromolecules, 30, 8, “Nanotribology: Friction, Wear and Lubrication at 1997, 2482-2494. the Atomic Scale,” Nature, 374, 6523, 1995, [15] Israelachvili, J., “Thin Film Studies Using 607-616. Multiple Beam Interferometry,” J. Colloid [4] Israelachvili, J. and Berman, A. D., Surface Forces Interface Sci., 44, 2, 1973, 259-272. and Microrheology of molecularly thin liquid [16] Yamada, S., “General Shear Thinning Dynamics films, CRC Handbook of Micro/Nanotribology, of Confined Fluids,” Tribol. Lett., 13, 3, 2002, 2nd ed., ed. Bhushan, B., CRC Press, Boca Raton, 167-171. FL, 1999, 371. [17] Raviv, U., Tadmor, R. and Klein, J., “Shear and [5] Cai, L., Peanasky, J. and Granick, S., Frictional Interactions between Adsorbed Polymer “Nanorheology of Polymers,” Trends Polym. Sci., Layers in a Good Solvent,” J. Phys. Chem. B, 105, 4, 2, 1996, 47-51. 34, 2001, 8125-8134. [6] Yamada, S., Nakamura, G. and Amiya, T., “Shear [18] Tadmor, R., Chen, N. and Israelachvili, J., Properties for Thin Films of Star and Linear “Normal and Shear Forces between Mica and Polymer Melts,” Langmuir, 17, 5, 2001, Model Membrane Surfaces with Adsorbed 1693-1699. Hyaluronan,” Macromolecules, 36, 25, 2003, [7] Israelachvili, J. N., Intermolecular and Surface 9519-9526. Forces, 2nd ed., Academic Press, London, 1991, [19] Yu, C., Evmenenko, E., Kmetko, J. and Dutta, P., 288-311. “Effects of Shear Flow on Interfacial Ordering in [8] Yamada, S., “Layering Transitions and Tribology Liquids: X-ray Scattering Studies,” Langmuir, 19, of Molecularly Thin Films of 23, 2003, 9558-9561. Poly(dimethylsiloxane),” Langmuir, 19, 18, 2003, [20] Schmatko, T., Hervet, H. and Leger, L., “Friction 7399-7405. and Slip at Simple Fluid-Solid Interfaces: The [9] Granick, S., Lin, Z. and Bae, S. C., “Molecules Roles of the Molecular Shape and the Squeezed and Stroked,” Nature, 425, 6957, 2003, Solid-Liquid Interaction,” Phys. Rev. Lett., 94, 24, 467-468. 2005, 244501. [10] Fetters, L. J., Lohse, D. J., Milner, S. T. and [21] Evmenenko, G., Dugan, S. W., Kmetko, J. and Graessley, W. W., “Packing Length Influence in Dutta, P., “Molecular Ordering in Thin Liquid Linear Polymer Melts on the Entanglement, Films of Polydimethylsiloxanes,” Langmuir, 17, Critical, and Reptation Molecular Weights,” 13, 2001, 4021-4024. Macromolecules, 32, 20, 1999, 6847-6851. [22] Evmenenko, G., Mo, H., Kewalramani, S. and [11] Horn, R. G. and Israelachvili, J. N., “Molecular Dutta, P., “Conformational Rearrangements in Organization and Viscosity of a Thin Film of Interfacial Region of Polydimethylsiloxane Melt Molten Polymer between Two Surfaces as Probed Films,” Polymer, 47, 3, 2006, 878-882. by Force Measurements,” Macromolecules, 21, 9, [23] Maeda, N., Chen, N., Tirrell, M. and Israelachvili, 1988, 2836-2841. J., “Adhesion and Friction Mechanisms of [12] Gee, M. L. and Israelachvili, J. N., “Interactions of Polymer-on-Polymer Surfaces,” Science, 297, Surfactant Monolayers across Hydrocarbon 5580, 2002, 379-382. Liquids,” J. Chem. Soc. Far. Trans., 86, 24, 1990, [24] Chen, N., Maeda, N., Tirrell, M. and Israelachvili, 4049-4058. J., “Adhesion and Friction of Polymer Surfaces: [13] Israelachvili, J. and McGuiggan, P. M., “Adhesion The Effect of Chain Ends,” Macromolecules, 38, 8, and Short-Range Forces Between Surfaces I - 2005, 3491-3503. New Apparatus for Surface Force Measurements,” J. Mater. Res., 5, 10, 1990, 2223-2231.

Japanese Society of Tribologists (http://www.tribology.jp/) Tribology Online Vol. 1, No. 2 (2006) / 33